Author: Uri Blass
Date: 11:37:07 05/16/03
Go up one level in this thread
On May 16, 2003 at 13:43:45, Rolf Tueschen wrote: >On May 16, 2003 at 12:42:54, Uri Blass wrote: > >>On May 16, 2003 at 11:33:18, Rolf Tueschen wrote: >> >>>On May 16, 2003 at 10:22:48, Peter Berger wrote: >>> >>>>On May 15, 2003 at 18:50:51, Rolf Tueschen wrote: >>>> >>>>>Oh, excuse me... Do you really want to challenge Huebner? >>>> >>>>I didn't mean to challenge Huebner. I challenged his analysis. >>>> >>>>>Either you are someone >>>>>else under pseudonym or you are impostering. And you want to explain that the >>>>>opening with gxf6 is a good move in the game? I must admit I didn't have the >>>>>time to look at it. >>>> >>>>Very convincing argument ;) >>>> >>>>>But Huebner surely had the time and IF gxf6 would be a >>>>>boomer he certainly wouldn't have given NxNf6 and NxN in his analysis. Just IMO >>>>>of course. But Huebner is allegedly only the best German player, ok, if we >>>>>forget some who came from elsewhere after 1989. And he certainly would not >>>>>oversee such a possibility that _you_ could find after an analysis with FRITZ. >>>>>Ok so far or do you want to claim some higher status, higher than Huebner? :) >>>>> >>>> >>>>I got taken away because of my interest in this game, else I would certainly not >>>>have started to discuss it with you. >>>>IMHO analysis is a product of time, effort and ability. It is very likely that I >>>>spent much more time and effort on analysis of the 16. g3 position than Huebner, >>>>which might make up for some of my defiencies when it is about abilities, >>> >>> >>>I doubt this. To give an example. The experience of a GM and former candidate >>>for the Wch title like Huebner, who is famous for his detailed analysis, is >>>described with the number of chunks he controls in his sleep. Some 15000 >>>perhaps. Such experience you can never equalize with hours or days of analyses. >>>Now if Huebner did NOT analyse your continuation it is because a) he knows that >>>it is good for White or b) he didn't want to discuss it because it might have >>>taken too much time (unpaid time). I would not even say that the fact that he >>>didn't analyse your move meant anything at all. Perhaps he is waiting for >>>contradicting ideas - who knows. I often saw this in his analyses for the CBM >>>journals. He's always prepared for interesting findings. >>> >>>I think people have a false comprehension of GM analyses. I doubt that GM always >>>want to give the absolute truth. What would that help? No opponent plays for the >>>absolute truth. Certainly no computer. So the main part of analysis is what the >>>opponent moves. >>> >>>If you think that corresponance chess is the key to real human chess or PC vs >>>human chess, then you are wrong IMO. Simply because human chess is something >>>else. A GM knows what you are intending and he will change the road if he sees >>>that he's going into a certain automatism. I think that people often >>>underestimate this typical strategy in chess when they argue with their computer >>>analyses. To me it seems as if chess is only starting when theory or >>>correaspondance analyses say unclear. Then it is ok for a GM, because no >>>computer program is already GM! So by force the GM must win the advantage. >> >>I can say by the same logic that unclear position is ok for the computer >>because no GM can calculate so many lines like a computer >> >>Speed give advantage to computers and the only question is if the advantage is >>bigger than other advantages of the GM's. >> >>The results of computers suggest that the speed often gives enough advantage to >>the machines. > > >Yes, you are absolutely right. But as I said more than a hundred times, often >the GM in such shows do NOT play against the machines. They play their normal >chess which is already sub-optimal! Because then happens what you say. Often it >comes to tactical positions where a machine makes less faults than a human >player. > >However no sponsor wants to see what a GM really could do to a machine & its >program. I mean, Hollywood was also not created to tell the truth about human >relationships but to tell us the dream we all dream in our daily life. That is >why we pay for it. Because reality is much more complicated than a Hollywood >romance. > >What a GM could really do with a machine?? Ahem, - - putty? I think that the solution is to change the conditions. For example you can let every player the right to take back a move limited number of times. You can decide that the time is 90 minutes per game+ 2 minutes per move when only 5 take backs are allowed during the game. A player in his turn can take back 2 plies and play a move and it is going to be considered as one takeback. Taking back 4 plies and playing a move is considered as taking back 2 moves. I believe that humans can earn more from the right to take back and an interesting question may be what is the minimal number of take back that can do computers less than IM strength. Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.